Predicting drug-target interaction based on bilateral local models using a decision tree-based hybrid support vector machine

Identifying the interaction between the drug and the target proteins plays a very important role in the drug discovery process. Because prediction experiments of this process are time consuming, costly and tedious, Computational prediction can be a good way to reduce the search space to examine the interaction between drug and target instead of using costly experiments. In this paper, a new solution based on known drug-target interactions based on bilateral local models is introduced. In this method, a hybrid support vector machine based on the decision tree is used to decide and optimize the two-class classification. Using this machine to manage data related to this application has performed well. The proposed method on four criteria datasets including enzymes (Es), ion channels (IC), G protein coupled receptors (GPCRs) and nuclear receptors (NRs), based on AUC, AUPR, ROC and running time has been evaluated. The results show an improvement in the performance of the proposed method.

Predicting drug-target interaction based on bilateral local models using a decision tree-based hybrid support vector machine Keywords:

Drug-target interaction , bilateral local model , Decision Tree , hybrid SVM